Verity rhazime June2008 public - oss.jishulink.comoss.jishulink.com/caenet/forums/upload/2013/10/09/191/41842301202818.pdfJun 10, 2008 · – The VerityTM structural stress definition

Center for Welded Structures Research

Fatigue Strength of Welded JointsVerityTM in Fe-safe

Radwan Hazime, PhDSafe Technology (US) Limited

June 3th, 2008

Outline1. The Verity Method

– Background and needs– Stress based existing methods and codes– The VerityTM structural stress definition– Formulation of the master S-N curve– Validations and applications– Concluding remarks

2. Implementation in fe-safe and application examples

3. Modeling Considerations

4. Demo of fe-safe and Verity

1. The Verity Method for Welded Joints

Rails, railway vehicles and

plant

Ship building

Fatigue of welded joints

Ships Civil engineering

Fatigue of welded joints Fatigue of welded joints

Power generation Car bodies and car assembly plant

Car bodies and car assembly plant

“This will save millions of dollars” - Ford

Truck, bus and off-highway vehicles


Truck, bus and off-highway vehicles

Caterpillar and John Deere have expressed strong interest. Caterpillar has more than 30 fe-safe™ licences

Pressure vessels


Pressure vesselsMedical equipment -

X-ray treatment and mobile MRI scanners

Fatigue of welded joints Issues: welded joints

1. Welding process induces highly localized heating and cooling

3. Randomly distributed discontinuities or defects

Planar discontinuitiesVolumetric discontinuities(slag inclusion, Porosity)

2. Material Heterogeneity

Buckling Distortion

Issues: welded joints

4. Residual stresses can be as high as material yield

5. Stress concentration at weld toe or notch

6. Stress concentration at weld ends

7. Estimation of local stresses in HAZ is complex, costly, time consuming and are sensitive to mesh sizes

8. Structures will be subjected to complex time varying loads

Mesh-Sensitivity in Stress Calculations for Welded Joints

• Stress singularity at sharp notches

1.0

2.0

3.0

4.0

0.0

Norm

alize

d St

ress

Element Size (Δl/t)

F/A

Peak stress at Weld Toe from FE Model

1.0

2.0

3.0

4.0

0.0

Norm

alize

d St

ress


F/A


• Mesh-sensitivity in stress calculations

• Existing Codes/Standards: based on nominal stress –the distance from the weld toe is very subjective

BS7608 Joint Classification - Currently Used by Various Industries

Weld Classes and S-N Curves Used by IIW, Eurocodes, AWS, AASHTO, API, etc

Based on nominal stress – choice of reference distance is subjectiveDifferent S-N curves for each type of weld.

B C F F2

Mesh-Sensitivity in Stress Calculations for Welded Joints

• Stress singularity at sharp notches

1.0

2.0

3.0

4.0

0.0

Norm

alize

d St

ress


F/A


1.0

2.0

3.0

4.0

0.0

Norm

alize

d St

ress


F/A


• Mesh-sensitivity in stress calculations

• Existing Codes/Standards: based on nominal stress –the distance from the weld toe is very subjective

σtF

σtF

σσtF

0.4t1t

• … or on extrapolated hot-spot stress HSS

Extrapolated hot-spot stress (HSS)

σtF

σtF

σσtF

0.4t1t

Should it be 0.4t and 1t ?0.5t and 1.5t ?

Objective: to define a weld toe stress that characterises the fatigue life of the weld – and therefore a single S-N curve for all welds

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

0 5 10 15 20 25 30

Distance from Weld Toe

Stress

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

.5t/1

.5t

.4t/1

.0t

0.5t

Ext

rap o

lat e

dst

ress

es

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

.5t/1

.5t

.4t/1

.0t

0.5t

Ext

rap o

lat e

dst

ress

es

SCF

ExtrapolationProcedures


0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

0 5 10 15 20 25 30

Distance from Weld Toe

Stress

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

.5t/1

.5t

.4t/1

.0t

0.5t

Ext

rap o

lat e

dst

ress

es

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

.5t/1

.5t

.4t/1

.0t

0.5t

Ext

rap o

lat e

dst

ress

es

SCF

ExtrapolationProcedures

ExperimentShell4Shell4Shell8Shell8Shell4(css)Shell4Shell8w1Solid20wSolidpw2Solid20w4Solid8w4Solid8w2Solid20w(f)


Extrapolated HSS is very mesh-sensitive and very sensitive to extrapolation method

Requirements for a FE Based Stress Parameter Definition for Fatigue Evaluation

• Consistency in stress determination:– Mesh-insensitive– Robust for complex structures – always get the same

answer

• A single S-N curve should apply to:– different joint geometries– different loading modes– different plate thicknesses

Fatigue of welded joints –Verity™…. Welded joint behavior




e



Fatigue of welded joints –Verity™ in fe-safe™

• Developed and patented by the Battelle Institute

- Licensed exclusively to Safe Technology - world-wide agreement in place

• Dr Pingsha Dong has received many awards -Society of Automotive Engineers Henry Ford II medal - Time Magazine 2005 maths innovator

Verity Basics

Displacement based FE procedures:

• Nodal forces and displacements are most reliable solution quantities• Equilibrium conditions are only guaranteed in terms of nodal forces

at nodes, but not in terms of stresses

y’x’

N1

N2

N3NiE1

E2

E3Ei

WeldNodes at Weld Toe

Displacement based FE procedures:• Nodal forces and displacements are most reliable solution quantities• Equilibrium conditions are only guaranteed in terms of nodal forces

at nodes, but not in terms of stresses

The equilibrium-equivalent structural stresses can be extracted using:• Balanced nodal forces

“NLOAD” in ANSYS“NFORC” in Abaqus“GPFORCE” in Nastran

• “Work equivalent” based mapping from nodal force/moments to line force/moments

FEA Numerical Implementation Automated Procedures for Shell/Plate Models: Transforming Nodal Forces/Moments to Line Force and Moments

N1

y’x’

N2

N3NiE1

E2

E3Ei

WeldNode at Weld Toe

x

y

z

⎪⎪⎪⎪

⎭

⎪⎪⎪⎪

⎬

⎫

⎪⎪⎪⎪

⎩

⎪⎪⎪⎪

⎨

⎧

⎥⎥⎥⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢⎢⎢⎢

⎣

⎡

+

+=

⎪⎪⎪⎪

⎭

⎪⎪⎪⎪

⎬

⎫

⎪⎪⎪⎪

⎩

⎪⎪⎪⎪

⎨

⎧

nn f

fff

llll

llll

ll

F

FFF

.

.

......0063

)(6

0

063

)(6

0063

.

.3

2

1

3322

2211

11

3

2

1

Coordinate rotations and solving simultaneous equations:

Line forces along the weld line - work equivalent argument

⎩⎨⎧

=⎭⎬⎫

forces lineby donework

forces GP

by donework

Mesh Independent Structural Stress Evaluation – line forces

F1,u1

1 2 3

F2,u2 F3,u3

∫∫ +=++21 L

0

''L

0332211 dl u f dlu f u F u F u F

1 32

f1

f2

L1 L2

f3f,u

dl

f’,u’ 22112211

fN fN fuN uN u

+=+=

3221'

3221'

fN fN f

uN uN u

+=

+=

⎪⎭

⎪⎬

⎫

⎪⎩

⎪⎨

⎧

⎥⎥⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢⎢⎢

⎣

⎡

+=

⎪⎭

⎪⎬

⎫

⎪⎩

⎪⎨

⎧

3

2

1

22

2211

11

3

2

1

fff

3L

6L0

6L

3L L

6L

06

L3

L

FFF

GP forces line forces

Line forces

⎪⎪⎪⎪

⎭

⎪⎪⎪⎪

⎬

⎫

⎪⎪⎪⎪

⎩

⎪⎪⎪⎪

⎨

⎧

⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥

⎦

⎤

⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢

⎣

⎡

+

+

=

⎪⎪⎪⎪

⎭

⎪⎪⎪⎪

⎬

⎫

⎪⎪⎪⎪

⎩

⎪⎪⎪⎪

⎨

⎧

+++

=

⎪⎪⎪⎪

⎭

⎪⎪⎪⎪

⎬

⎫

⎪⎪⎪⎪

⎩

⎪⎪⎪⎪

⎨

⎧

nn

nnn f

fff

l

l

llll

llll

ll

F

FFFFF

F

F

FFF

.

.

6..............

..................

.........6

00

......63

)(6

0

......063

)(6

......0063

.

...

3

2

1

3

3322

2211

11

)(

)4(3

)3(3

)3(2

)2(2

)1(2

)1(1

3

2

1

Element forcesGP forces

System of equations for a general curved weld line

•Mesh independent•Numerically robust•Consistent line forces

Similarly, line moments along the weld line - work equivalent argument

⎩⎨⎧

=⎭⎬⎫

moments lineby donework

moments GP

by donework

Mesh Independent Structural Stress Evaluation-stresses•Calculate structural stress components fromline forces and moments

2bms tm6

tf

stress structural Total

+=+=⎭⎬⎫

σσσ

Bending structural stress

Line moment

Thickness of sheetMembrane structural stress

•Work equivalent arguments are coupled with a special virtual nodemethod to capture the stress concentration at weld ends

X’, f

Y’,m

Line force

0.25tx0.25t 0.25tx0.25t

(a) Tubular T-JointHot Spot

Chord

BraceHot SpotHot Spot

Chord

Brace

0.5tx5t 1tx1t 2tx2t

Saddle

0.5tx5t 1tx1t 2tx2t 2tx2t

Saddle

Mesh Insensitivity:A Tubular Joint (Zerbst et al, 02)

2

4

6

8

10

12

0 30 60 90Angle from Saddle Point (Deg.)

SCF 2tx2t

1tx1t0.5tx0.5t0.25tx0.25t

(c) Structural stress SCF results

SaddleCrown

Peak SS

Structural stress is mesh-insensitive

0.5tx5t 0.5tx5t 0.5tx5t 0.5tx5t

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

0 30 60 90 120 150

Distance from top of attachment, mmSt

ruct

ural

Str

ess,

MPa

shell-0.5tx0.5trshell-1.0tx1.0trshell-2.0tx2.0tr

Mesh-Insensitive SS Demonstration – Gussets on Plate Edge (FPSO Detail 5)

0.25tx0.25t 0.5tx0.5t

1.0tx1.0t 2.0tx2.0t

Weld End (Peak Stress)

Structural stress is mesh-insensitive

σtF

σtF

σσtF

0.4t1t

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

0 5 10 15 20 25 30

stre

sses

at R

OP

's


0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

.5t/1

.5t

.4t/1

.0t

0.5t

Ext

rapo

late

d st

ress

es

Distance from Weld ToeExtrapo lationProcedure s

SCF

Attachment

Base Plate

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

0 5 10 15 20 25 30

stre

sses

at R

OP

's


0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

.5t/1

.5t

.4t/1

.0t

0.5t

Ext

rapo

late

d st

ress

es

Distance from Weld ToeExtrapo lationProcedure s

SCF

Attachment

Base Plate

FPSO Phase 1 Results (Fricke,01)

Extrapolated HSS is mesh-sensitive

Focus on rat hole end

Bracket

Web Frame

Side Shell

Longitudinal Stiffener Web

Web Frame Stiffener Web

Mesh Insensitivity: Recent Comparative Study on HSS and Structural Stress Methods (B. Healy)

Focus on rat hole endFocus on rat hole end

0

1000

2000

3000

4000

5000

abaqus-8r abaqus-4 abaqus-4r nastran-8r nastran-4

2t

t

0.5t

0.25t0.125t

Structural Stress Method

0

1000

2000

3000

4000

5000

abaqus-8r abaqus-4 abaqus-4r nas tran-8r nas tran-4

2t

t

0.5t0.25t

0.125t

HSS (.5t/1.5t)HSS (.4t/1t)

0

1000

2000

3000

4000

5000

6000

0

1000

2000

3000

4000

5000

6000

Focus on rat hole end

Bracket

Web Frame

Side Shell

Longitudinal Stiffener Web

Web Frame Stiffener Web

Mesh Insensitivity: Recent Comparative Study on HSS and Structural Stress Methods (B. Healy)

Stress Intensity Factor Estimation Using Structural Stresses

2c

a tr

General 3D Welded Joints

The structural stress at the weld toe in mesh-insensitive, but is that enough – what about crack growth / specimen compliance effects ?

Crack growth / compliance

t

σs

σb

σb

(a) Membrane Dominated

(b) Bending Dominated

Smalls

brσσ

=

Larges

brσσ

=

(a) Remote Loading Mode Effects

(b) Thickness Effects

t

t

F

8t

25t

1

1.1

1.2

1.3

1.4

1.5

1.6

No

rma

lize

d S

tru

ctu

ral

Str

ess

Dimensions are Proportional for 3 Joints

t 2t 3t

Stress Intensity Factor Estimation Using Structural Stresses

2c

a tr

General 3D Welded Joints

[ ] where

)(

:Cracks Edge

r

bms

bmbms ffftKσσσ

σσ+=

−−= ta

σbσm

ta

σbσm

Newman and Raju or alet Shiratori either from Y and Y where

2-

:Cracks Elliptical

10

b 10 2)( YQaY

QaK bs

πσπσσ +=

Weld

t σx (y)τ(y)

mσ bσ

Weld

t τm

Structural Stress: Equilibrium Equivalent

Notch Stress: Self -equilibrating

Weld

t

FE Model

Mesh sensitive

•Uses far field nominal stress •OK for a/t >0.1• Long crack

•Using local stress concentration (notch)• a/t< 0.1 • Short crack

a= crack size at weld tow

The VerityTM Structural Stress Definition Local Analysis – Mode I Crack Growth Stress Intensity Factor Approach

Complex geometry &loading in Global analysis

Simple geometry &loading

t

σm σb

a

Problem definition: A finite width plate with a surface crack “a”subjected to remote loads

Structural stress components are the remote loads (tension and bending) for local analysis

⎟⎠⎞

⎜⎝⎛Δ=Δ

⎭⎬⎫

taftK

load tensile todue RangeFactorIntensity Stress

mmm σ

functions compliance bending and membrane are f andf taftK

load bending todue RangeFactorIntensity Stress

bm

bbb ⎟⎠⎞

⎜⎝⎛Δ=Δ

⎭⎬⎫

σ

bm K K Kload combined todue RangeFactorIntensity Stress

Δ+Δ=Δ⎭⎬⎫

s

b

bm

br ratio, bending definingBy σσ

σσσ

ΔΔ

=Δ+Δ

Δ=

⎥⎦

⎤⎢⎣

⎡⎥⎦

⎤⎢⎣

⎡⎟⎠⎞

⎜⎝⎛−⎟

⎠⎞

⎜⎝⎛−⎟

⎠⎞

⎜⎝⎛Δ=Δ

taf

tafr

taftK bmmsσ

The stress intensity factor range can be written as

f= compliance functions

Factorion MagnificatIntensity Stress is M

andconstant a is C ere whK)()C(MdNda

kn

mnkn Δ=

Two-stage crack growth law:

) and s thickneson through (basedKeffects)notch localwith (KM

bm

Notchkn σσ=

Life prediction in cycles to final failure

∫∫=

=

=

= Δ=

Δ=

1a/t

0a/tmn

kn

aa

0amn

kn K)()C(Mtd(a/t)

K)()C(MdaN

f

m= Paris law exponentn= (taken as = 2) unifies short crack growth rate with long crack growth.

Master SN curve

∫∫=

=

=

= Δ=

Δ=

1a/t

0a/tmn

kn

aa

0amn

kn K)()C(Mtd(a/t)

K)()C(MdaN

f

∫=

=

−

Δ=

faa

0a

2/1

)/()/(1N tadtaFC

tms

m

σ

)(1N2/)2(

rIC

tms

m

σΔ=

−

)((N)(C)2/)2(

rItms

m

σΔ=

−

m

s

mmm rItC /1

2/)2(/11/m )]([N

σΔ=

−

mmmsm

rItC

/12/)2(/11/m-

)]([N

−

− Δ=σ

Taking C, t, and Δσs out of the integral,

Equivalent Structural Stress

after substituting for ΔK and Mkn,

m-1

m-1

S N C SStress StructuralEquivalent

=Δ⎭⎬⎫

where,

m1

2mm-2

SS

I(r) t S σΔ=Δ

Master S-N curvem=Paris law exponent

This equivalent structural parameter includes the effects of thickness, geometry, bending ratio and test conditions I(r).

Structural stress

Two-stage crack growth law:

1a/t

0a/tm

bmmn

kn ta

fta

frta

f)M(

d(a/t)I(r) ∫

=

=⎥⎦

⎤⎢⎣

⎡⎟⎠

⎞⎜⎝

⎛⎟⎠⎞

⎜⎝⎛−⎟

⎠⎞

⎜⎝⎛−⎟

⎠⎞

⎜⎝⎛

=

I(r) is a dimensionless function of r and is given by,

I(r) function can be expressed in actual test loading conditions (e.g. displacement controlled condition or load controlled condition).

Two-stage crack growth law: The Fatigue Governing Parameter: Equivalent Structural Stress Parameter Δss

Modify the structural stress for effects of

s

b

sm

brσσ

σσσ

=+

=- loading mode

- Thickness t

… to produce an equivalent structural stress

“Loading Mode Effect”

mmm

ss

rItS 1

22

)(⋅

Δ=Δ −

σ

“Thickness Effect”

“Stress Concentration Effect” I(r)^(1/m), m=3.6

1

1.1

1.2

1.3

1.4

1.5

1.6

0 0.2 0.4 0.6 0.8 1Bending Ratio (r)

I(r)^

(1/m

)

Load Controlled Disp Controlled

s

b

sm

brσσ

σσσ

=+

=

mrI1

)( as a function of r

The Fatigue Governing Parameter: Equivalent Structural Stress Parameter Δss

Notch Stress Intensity Magnification Factor (Mkn) Estimation

Sym.

Notched specimen

Remotebending

Mkn is dominant approximately within a/t < 0.1

Mkn=


mmm

ss

rItS 1

22

)(⋅

Δ=Δ −

σ


“Stress Concentration Effect”

Equivalent structural stress

s

b

sm

brσσ

σσσ

=+

=…loading mode

…thickness

…is calculated at the weld toe

How well does it correlate with test results ?

The Fatigue Governing Parameter: Equivalent Structural Stress Parameter Δss

pp

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

0 30 60 90 120 150

Distance from Top Weld Toe on Attachement, mm

Nor

mal

ized

Str

uctu

ral S

tres

s

0.5tx0.5t1tx1t2tx2t4tx4t

Top Weld Toe

Bottom Weld Toe(b) Comparison of structural stress distributions

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

0 30 60 90 120 150

Distance from Top Weld Toe on Attachement, mm

Nor

mal

ized

Str

uctu

ral S

tres

s

0.5tx0.5t1tx1t2tx2t4tx4t

Top Weld Toe

Bottom Weld Toe(b) Comparison of structural stress distributions

Virtual Node Method – Stress Concentration at Weld Ends

Before Virtual node treatmentAfter Virtual node treatment

Stress concentration with different mesh sizes

A Plate to Frame Joint

Correlation: All Pipe and Vessel Weld S-N Data (~500 tests) – ASME Div 2 Rewrite JIP)

1.E+01

1.E+02

1.E+03

1.E+04

1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08

Life

Equi

vale

nt S

truct

ural

Stre

ss R

ange

, MP

a

Equivalent Structural Stress Range

mmm

ss

rItS 1

22

)(⋅

Δ=Δ −

σ

1.E+01

1.E+02

1.E+03

1.E+04

1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08

Life

Nor

min

al S

tress

Ran

ge, M

Pa

ASME Mean

ASME III Design

Markl’s Equation(Mean Line for i =1)

BS5500 Design(Smooth ground butt welds)

Nominal Stress Range

1.E+01

1.E+02

1.E+03

1.E+04

1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08

Life

Stru

ctur

al S

tress

Ran

ge, M

Pa

ASME Mean

ASME III Design

Markl’s Equation(Mean Line for i =1)

BS5500 Design(Smooth ground butt welds)

Structural Stress Range

t

σs

σb

σb

t

t

F

8t

25t

1.E+02

1.E+03

1.E+04

1.E+04 1.E+05 1.E+06 1.E+07

Life

Equ

ival

ent S

truct

ural

Stre

ss R

ange

, MPa

AT122 AT140 AT180 AT222AT240 AT280 Bell Joint G'Joint D Detail_3(Fricke) Joint F joint F(rorup)Joint-Cb(Booth) Joint-Cb(Pook) AC110 AC122AC140W AC140N AC180 AC210AC222 AC240 AC280 AC310AC340 AC380 AC422 AC440Joint C Joint B 13/10/8 AW 50/50/16 AW50/50/16 AW (DW) 100/50/16 AW 100/50/16 AW (QT) Joint EGurney -LW2 HHI_3 9mm-w25 9mm-w509mm-w100 9mm-w160 20mm-w25 20mm-w5020mm-w100 20mm-w160 40mm-w25 40mm-w5040mm-w100

mmm

ss

rItS 1

22

)(⋅

Δ=Δ −

σ

1.E+02

1.E+03

1.E+04

1.E+04 1.E+05 1.E+06 1.E+07

Life

Equ

ival

ent S

truct

ural

Stre

ss R

ange

, MPa


mmm

ss

rItS 1

22

)(⋅

Δ=Δ −

σ

(b)

1.E+02

1.E+03

1.E+04

1.E+04 1.E+05 1.E+06 1.E+07

Life

Equ

ival

ent S

truct

ural

Stre

ss R

ange

, MPa


mmm

ss

rItS 1

22

)(⋅

Δ=Δ −

σ

1.E+02

1.E+03

1.E+04

1.E+04 1.E+05 1.E+06 1.E+07

Life

Equ

ival

ent S

truct

ural

Stre

ss R

ange

, MPa


mmm

ss

rItS 1

22

)(⋅

Δ=Δ −

σ

(b)t

Joint Gb (t=20mm)

t

Joint B(t=12.7mm), Joint B(Kihl)(6.35mm),13/10/8AW(13mm), 50/50/16AW(50mm), 50/50/16AW(DW)(50mm),100/50/16AW(100mm),100/50/16AW(QT Steel)(100mm)

t

Joint C(t=12.7mm)

t

Joint D(t=12.7mm)

t

Joint F (t=12.7mm), Joint F(Rorup)(12.5mm)

t

Bell (t=16mm)

Double Edge Gusset (90mm)

t

Joint G’ (t=12.7mm)

t

Joint-Cb(Booth)(t=38mm), Joint-Cb(Pook)(38mm)

t

Joint E (t=12.7mm)

t

t = 5-80mm

t

Joint Gb (t=20mm)

t

Joint B(t=12.7mm), Joint B(Kihl)(6.35mm),13/10/8AW(13mm), 50/50/16AW(50mm), 50/50/16AW(DW)(50mm),100/50/16AW(100mm),100/50/16AW(QT Steel)(100mm)

t

Joint C(t=12.7mm)

t

Joint D(t=12.7mm)

t

Joint F (t=12.7mm), Joint F(Rorup)(12.5mm)

t

Bell (t=16mm)

Double Edge Gusset (90mm)

t

Joint G’ (t=12.7mm)

t

Joint-Cb(Booth)(t=38mm), Joint-Cb(Pook)(38mm)

t

Joint E (t=12.7mm)

t

t = 5-80mm

Plate Joints with various thickness

Correlation: All Literature Data (> 800 Tests) – Load Controlled

Correlation for spot welds

Thus, we have a fatigue parameter,

•Insensitive to mesh

•Takes into account stress concentration, thickness, and loading mode effects

•A single SN curve for all welded joints of a given class of material (Steel, AL, Titanium, ..)


mmm

ss

rItS 1

22

)(⋅

Δ=Δ −

σ


“Stress Concentration Effect”

Verity: Equivalent Structural Stress Parameter Δss

2. Implementation in fe-safe and application

examples

B.C. & loads

Loading for superpositionStress &

nodal forcesDesign FE model

FEAABAQUS

NASTRANANSYSPro/E

I-DEAS

FE SafeStructural Stress

Life ContourRedesign

Verity for welds

Implementation in fe-safeTM

Implementation in fe-safeTM for scaling and combining load cases

1. For each load case, the membrane and bending structural stresses are calculated

2. Membrane and bending structural stresses are scaled and combined separately

3. I(r)1/m is calculated for each point in time; r = |σb|/( |σb|+|σm|)

4. Membrane and bending structural stresses are added and te equivalent structural stresses are calculated for each point in time

5. The equivalent structural stresses history is cycle counted and the damage is added

Superimposed load histories

Loading

SignalSingle load history

Modal superimposition+

+

Sequencies of FEA solutions

SAE FD&E “Fatigue Challenge” Blind Life Prediction

• SAE FD&E issued a “fatigue prediction challenge”

• Actual test results were given after all participants presented their predicted lives

• See www.fatigue.org/weld

• The Verity method won “The Best Prediction:

Application - MIG welded T-box (Constant Amplitude)

Fine MeshCoarse mesh

Life Contour Plots - Mesh insensitivity

Test results (Initiation + Propagation)75,000Experimental (Deere)w. surf. Rough corr. Factor 0.65

‘text book’ guess based on FEAParticipant 5Not as fixed35,000

Fixed body condition51,000Participant 5Mesh 368,900Mesh 266,100Mesh 177,100Dong, Battelle -Verity90% penetration674,000

50% penetration211,000

10% penetration54,000

3D model, full penetration53,000Participant 3

Hand Calculations30,000Participant 2BS 7608 Class G33,529

BS 7608 Class W21,800Participant 1CommentNf (R=-1)Who

Verity results voted to be the best

Life Predictions from Weld Challenge ParticipantsA 2nd SAE Weld Challenge: Variable Amplitude Loading of Same Specimens

Weld end is much bigger in Challenge 2A

• The Verity method predicted the crack location and the fatigue life

Challenge 1 (2003)

Challenge 2A (2004)

Challenge 2A (2004)

Weld Representation at Weld Ends

Challenge 2A (2004)Model 1

Weld Representation at Weld Ends

Challenge 1 (2003)FF

Comparison of FE Models Used for Weld Challenge 1 (03) and Challenge 2A (04)

Identification of Critical Locations after Searching Two Weld Toe Lines

-300

-200

-100

0

100

200

300

-50 0 50 100 150 200 250

Distance from tube end, mm

Stru

ctur

al S

tress

, MPa

Challenge 1 (03)Challenge 2A (04)-Model 1Challenge 2A(04)-Model 2

-300

-200

-100

0

100

200

300

-50 0 50 100 150 200 250

Distance from tube end, mm

Stru

ctur

al S

tress

, MPa

Challenge 1 (03)Challenge 2A (04)-Model 1Challenge 2A-Model 2

2”x6” weld toe 4”x4” weld toe

F=4000 Ibs

Observations:• If the weld ends are big (modeled as posted in the website), weld end failure occurs on 4”x4”• if the weld ends are as small as those for Challenge 1, failure occurs at 2”X6” weld toe corner

2A

The method is the only method predicting both failure location and mean life correctly

SAE MIG Weld, MIG welded T-tube subjected to variable amplitude load

Load

(N)

TimeLocation of min life

Grapple skidder torque history (GSTH)

Life (blocks)Load Verity TEST

27.1 x GSTH 364 45019.2 x GSTH 1044 1) 1750

2) 2161

Weld end modeling

Concluding Remarks

mmm

ss

rItS 1

22

)(⋅

Δ=Δ −

σ

The Equivalent Structural Stress based Master S-N curve provides a single parameter description of

• Thickness (t)• Loading mode (r)• Stress concentration (Δss)

Validated by correlating S-N data from about 3500 fatigue tests from 1947 to presentWon SAE “Weld Challenge” twice in a row (2003 and 2004)Adopted by ASME Div 2 Rewrite (Design by Analysis)Implemented in fe-safe from Safe Technology to combine the analysis of welded and non-welded structures

Acknowledgement:Many thanks to Dr. Pingsha Dong of Battelle for supplying some of the slides

Documents

Verity rhazime June2008 public - oss.jishulink.comoss.jishulink.com/caenet/forums/upload/2013/10/09/191/41842301202818.pdfJun 10, 2008 · – The VerityTM structural stress definition